One page · Five domains × four levels · Designed for a 30-minute observation: mark behaviors as you see them, then transfer to this sheet afterward. Domain names and behaviors are the competency model’s, unchanged.
How to use it
Observe for 30 minutes with the behavior list (print version) ticking what you see. Ticks, not judgments — anchors are visible/audible events by design.
Score each domain 1–4 afterward using the anchors below. When a domain sits between two levels, the lower level wins unless the higher behavior is clearly present. Write one evidence line per domain — the tick you’d show the tutor.
One strength + one focus for the coaching conversation. Never deliver all five domains as verdicts — the cycle works on one focus move at a time.
Calibrate first. Scores only mean something after the calibration activity; new observers score T1, T3 and T5 and compare against anchors before observing live tutors.
Levels in one line each: 1 = absent or the failure mode dominates · 2 = present but generic or inconsistent · 3 = consistent, student mostly produces · 4 = consistent, specific, and student does the thinking — you could hand this session to a new tutor as a model.
Domain 1 · Relationships
Level
What you’d see and hear
4
Every student greeted by name within a minute; a specific reference to each student’s recorded work appears unprompted; attention visibly distributed — every student produces at least once; references name effort and process (“your distributive step got cleaner”), not just scores
3
All students greeted by name; at least one reference to previous work, specific enough to check against the tracker; most students addressed
2
Greetings happen but generic (“hey guys”); references vague (“let’s keep working on equations”); one student does most of the talking, tutor notices but doesn’t redistribute
1
Cold start into content; no names or names missed; one student carried, another invisible; no connection to any recorded work
Domain 2 · Session structure
Level
What you’d see and hear
4
All five blocks run inside their boxes (warm-up 5′ → review 5′ → guided practice 12′ → independent try 5′ → exit ticket 3′); visible timekeeping; exit ticket collected from every student including incomplete ones; review topic traced to ticket data out loud (“Tuesday’s tickets said…”)
3
All five blocks present; one block runs over but another still happens; exit ticket collected; review uses ticket data
2
Routine recognizable but blocks blur — guided practice swallows the hour, independent try shrinks; or the exit ticket is threatened (“we’ll do it next time”) and only intervention of the clock saves it
1
No routine: cold start, no review, no independent try, exit ticket skipped (this alone caps the domain at 1 — it breaks the data loop for the whole group; see T5 annotation 7)
Domain 3 · Questioning and errors
Level
What you’d see and hear
4
Every problem opens with a probing question before any explaining (L0); ladder climbs one rung per turn, hand on the rung; errors met with a question at the misconception’s spot — never the label (“what did the 2 multiply?”, never “you forgot to distribute”); 3–5 second pauses after questions (count them); after ~3 failures, a worked example of a similar problem (L4), then the original handed back; the student produces the key steps
3
Probing questions asked, though sometimes after a partial explanation; wait time present but short (~2s); ladder mostly one rung per turn; errors probed at least once before any correction
2
Questions asked but self-answered; wait time absent; ladder skipped (L0 → “here’s what you do”); errors corrected outright with the right step supplied — tutor performs the fix (T6 annotation 3)
1
No probing questions; session is explanation (T5 annotation 1); tutor completes every solution; misconception named and patched (“that’s your bug”)
Domain 4 · Feedback and motivation
Level
What you’d see and hear
4
Feedback names the work (“you moved the −4 and flipped its sign — that’s the step from last week”); after correct answers: brief confirm + ONE why-question + offer of a similar problem with a small variation; errors normalized warmly and fast (“that’s the trap this one sets”); one small goal set with the group
3
Specific feedback appears most times success is acknowledged; why-questions asked sometimes; a similar problem offered at least once; no generic-only praise
2
Praise frequent but generic (“good job” ×N — count it); why-questions rare; no similar-problem offers; praise occasionally rewards speed
1
No feedback distinguishable from noise, or praise is entirely person-level (“you’re so smart,” “fastest in the group”); errors treated as failures to hide
Domain 5 · Data use
Level
What you’d see and hear
4
Exit ticket given and collected (and it shows: items pre-chosen in the plan); tracker recorded same day with tags; next session’s review topic and guided-practice target visibly derived from the results; stuck student flagged after two consecutive low tickets — flag includes the tagged tickets
3
Ticket given and collected; tracker recorded same day or next morning; plan shows a data-informed choice; no flag where one was warranted
2
Ticket collected but results never read (“a folder nobody read”); tracker updated from memory days later; next plan identical regardless of results
1
No exit ticket (see Domain 2 level 1); no tracker; data loop nonexistent
Print version (behavior list for the 30 minutes)
Tick during observation; transfer to scores after.
D1 Greet each by name (first minute) ☐
D1 Specific reference to recorded work ☐ (who: ____)
D1 Every student produces at least once ☐
D2 Warm-up ≤5′ / Review ≤5′ / GP ≤12′ / IT ≤5′ / ET ≤3′
(actual times) ____ ____ ____ ____ ____
D2 Exit ticket collected (incl. incomplete) ☐
D2 Review topic named from ticket data ☐
D3 Probing question before explaining (×___) ☐
D3 One ladder rung per turn (rungs: ______) ☐
D3 Wait time 3–5s (counted: ___ s) ☐
D3 Error → question at the spot, not a label ☐ (the Q: ______)
D3 L4 worked example if 3 failures ☐ n/a ☐
D4 Specific feedback naming the work (×___) ☐
D4 Confirm + one why + similar problem ☐
D4 Error normalized ("that's the trap…") ☐
D5 Ticket items pre-chosen in the plan ☐
D5 Tracker recorded same day (verified) ☐
D5 Next plan visibly derived from results ☐
D5 Flag raised where warranted ☐ n/a ☐
Scores and their use
Domain
1
2
3
4
Evidence line
1 Relationships
2 Structure
3 Questions/errors
4 Feedback/motivation
5 Data use
Program aggregates: the mean across observations (fidelity metric, target ≥3.0), growth between each tutor’s observations 1 and 3 (target +0.5), and the per-domain pattern (a tutor can be 4-3-2-4-3 — the pattern is the coaching plan). See measurement plan.
Fictional program, real scoring discipline. The anchors were written to be checkable — every level-4 behavior is a tick on the print sheet. When you can’t tick it, you can’t score it, and the debrief says so.